Background: Although the construct of depression has been subjected to numerous factor analytic studies and phenomenological subtypes of clinical relevance have been delineated, this is not the case for mania. The few available studies have reported at least two factors, which consist of euphoric versus dysphoric-hostile subtypes. Our objective was to replicate and further enrich this literature. Methods: In the EPIMAN French National Study we systematically evaluated 104 DSM-IV hospitalized manic patients in four university centers in different regions of France. Psychiatrists completed the Beigel-Murphy Manic State Rating Scale (MSRS), as well as the HAM-D-17, affective temperament scales, and the GAF Axis V from DSM-IV. Categorization of patients into pure versus dysphoric mania was made on the basis of clinical diagnosis, independent from psychometric measures. Results: On principal component analysis of the MSRS, three factors explained the largest variance: a global manic (23.3% variance), paranoid-hostile (14.8% variance), and psychotic (9.1% variance). After varimax rotation, we obtained seven independent factors: F1 Disinhibition-instability, F2 Paranoia-hostility, F3 Deficit, F4 Grandiosity-psychosis, F5 Elation-euphoria, F6 Depression, and F7 (Hyper)sexuality. We could not demonstrate significant correlations between the individual factors and impaired functioning on GAF. However, depressive and, to some extent, cyclothymic temperaments correlated with F6 Depression. Finally, intergroup comparisons between pure versus dysphoric mania diagnosed clinically showed high levels of F3 Deficit and F5 Elation in the pure, and of F6 Depression in dysphoric, mania; F2 Paranoia-hostility did not discriminate these two clinical forms of mania. Limitations: Although the present analyses on the Beigel-Murphy represent the largest sample studied to date, they are still underpowered and do not guarantee a stable factorial structure. Our findings are cross-sectional and require prospective validation. Conclusions: Our data suggest that 'dysphoria' as used in the literature to qualify mania is insufficiently precise, and is best further specified as 'depressive' versus 'irritable.' Moreover, our data extend the rich multidimensional phenomenology of mania beyond the existing literature: we submit that disinhibition-instability (a core 'activation' component) can, on the one hand, be associated with distinct emotional presentations (euphoric, depressive, or irritable-hostile), as well as psychotic and deficit symptomatology on the other. (C) 2002 Elsevier Science B.V. All rights reserved.
